James M. Krueger, Ph.D. 

Common knowledge of sleep is replete with "wives’ tales". Two perhaps are told to almost everyone by a loving parent or grandparent. Failure to obtain sufficient sleep renders one vulnerable to disease; and conversely if sick, sleep will help you recover. Despite the fact that even Hippocrates dispensed similar advice and that they seem so common sensible these postulates received scant scientific attention until recently. Within the past 10 years the questions of whether sleep loss affects host defenses and whether sleep is altered during infection and helps in recuperative processes have begun to be addressed. We review herein some of these recent findings. We conclude that induced sleep loss affects host defense and that it is likely beneficial for recovery from infectious diseases.

Sleep loss and host defense

Among the possible functions proposed for sleep, an immune system function had not garnered much attention until recently. This may have been due, in part, to a scientific emphasis on parsimony, which has engendered simple explanations for biological events, and therefore favored the expectation that the deprivation of a such a time-consuming state as sleep should produce a specific, obvious, and dramatic deficiency, much like an organ extirpation. In the last several years our thinking has changed. We no longer expect to find one big functional deficiency induced by sleep deprivation. Rather, sleep deprivation appears to induce a series of homeostatic responses to the missing requirement of sleep aimed at protecting the integrity of the organism. Physiological changes resulting from sleep deprivation develop slowly, and many days are required for clinical signs to become apparent. This is similar to what is expected during food or water deprivation. These basic deprivations of bodily requirements do not produce malfunction in the short-term, but they are lethal in the long-term, and yet the physical requirement is normally fulfilled on a daily basis.

Anecdotal evidence suggests an increase in upper respiratory tract infections during short-term sleep deprivation. Yet, experimental studies in humans have been unable to show an increased occurrence of illness related to sleep loss. However, the subjects in these studies typically are well rested beforehand and kept comfortable and fairly isolated during the study to avoid potential confounding variables. The experimental environment, then, allows protection from exposure to pathogens. The ramifications of sleep deprivation are expected to be much greater for an individual with compromised health, compared with a normal experimental subject sleep deprived on an acute basis. Indeed, sleep disruption is a risk factor for increased mortality, and it is believed to exacerbate disease processes.

In contrast to human work, animals studies can more directly address sleep loss-induced changes in host defenses. Brown and colleagues in Australia studied responses to a primary infectious challenge during short-term sleep deprivation in mice (Reviewed 1). Mice were inoculated twice, 7 days apart, with influenza virus administered into the gastrointestinal tract. A week later the mice were given live virus intranasally and half of the mice were sleep-deprived for seven hours, and half were allowed to sleep normally. (A mouse typically sleeps 11 or more hours per day.) Three days after the challenge, virus clearance from the lung was complete in mice allowed to sleep normally. By contrast, three days post challenge, virus clearance from the lung was incomplete in mice deprived of sleep. These findings suggest respiratory tract immune suppression due to sleep deprivation. Given the potential importance of these findings to health (e.g., influenza virus kills many more people each year than does HIV) these experiments should be replicated and extended to humans.

Brown and colleagues also investigated antibody production in sleep-deprived rats and found it to be suppressed (Reviewed 1). Rats were twice injected subcutaneously two weeks apart with a commonly used antigen, sheep red blood cells. After the second administration of antigen, the rats were given one of three agents, i.e., saline as control; the cytokine, interleukin-1; or the bacterial cell wall component, muramyl dipeptide. The rats then were either sleep deprived for eight hours or allowed to sleep normally. Sleep deprivation reduced antibody levels in the serum. The suppressant effects of sleep deprivation on antibody production responses to antigenic challenge could be eliminated if the sleep deprivation was combined with administration of interleukin-1 or muramyl dipeptide, factors that might have stimulated the immune system and thereby overridden the suppressant effects of sleep deprivation.

In short-term sleep deprivation studies in humans, uncomplicated by physical endurance regimens or other experimental stressors (e.g., military battle simulations used to gauge performance), seems to evoke the first line of host defense (Reviewed 1). A rise in the number of circulating white blood cells (WBCs) to concentrations above normal is commonly found. During two nights and three days of sleep deprivation the rise in circulating WBCs is followed by an increase in natural killer cell numbers and their cytotoxicity, shown by Dinges and colleagues at the Univerity of Pennsylvania. Despite conventional wisdom, activation of the hypothalamic-pituitary-adrenal (HPA) axis indicative of emotional distress and immunosuppression is not found in sleep deprivation studies that control for extraneous variables. At least five independent laboratories have demonstrated that the ability of circulating immunocytes to produce immune response modifiers, such as interleukin-1 (IL-1), tumor necrosis factor (TNF), interferon gamma or interleukin-6, is enhanced by sleep deprivation (5). These changes hint at an increased communication among and/or activation of other cells. Nevertheless the functional implications of an increase in circulating host surveillance mechanisms are unclear. Determination of their maladaptive versus regulatory nature awaits further study.

Short-term sleep deprivation of eight hours in rats in a yoked-control paradigm also is associated with an enhancement of host defenses as evidenced by a suppression of experimentally-induced tumor growth, as shown by Bergmann and colleagues at the University of Chicago (Reviewed 1). In this case, sleep deprivation appears to have an adaptive advantage in combatting subdermal foreign agents.

Long-term sleep deprivation in rats produces a reliable series of physical change of a syndromic nature. Most of the sleep-deprivation survival period in rats is marked by progressive development of a marked negative energy balance (manifested by dramatic increases in food consumption and loss of body weight), a progressive decline in plasma thyroid hormone concentrations very low levels (due to a change in central regulation), and sympathetic activation without HPA axis overactivation (Reviewed 2). These changes culminate in lethal sepsis, manifested by bloodstream infection that is antecedent to a cachectic-like moribund state comprising the last hours of days of survival. The septicemia is not accompanied by fever or marked tissue inflammatory reactions, indicating multiple alterations in host defense mechanisms have occurred (3).

At the onset of septicemia, sleep-deprived rats remain motorically active and many are hyperphagic. Therefore, the breakdown in host defense processes is antecedent to severe morbidity. The signs that follow are consistent with the expected outcome of cytokine and endotoxin-mediated pathology characteristic of lethal sepsis and shock. The bacteria that have been cultured in the bloodstream of sleep-deprived rats are mostly facultative anaerobes, indigenous to the host and to the environment. They are the same virulent microbes that cause risk to patients with suppressed immune systems; e.g., Pseudomonas aeruginosa, Streptococcus agalactiae, Klebsiella pneumoniae. Detection of virulent opportunistic microbes in the blood of sleep-deprived rats is evidence of a functional impairment. Once host defenses are breached, deleterious effects could be triggered by a number of agents other than bacteria, which are known to cause similar metabolic derangements, e.g., endogenous cytokines, viruses, and fungi.

The breakdown in host defense and the likelihood of immune compromise in sleep-deprived rats is the first physiological consequence of sleep deprivation identified that has obvious clinical significance. This finding provides a promising course of investigation to determine the mediation of earlier sleep deprivation signs. Ongoing studies are aimed at determining the timing and location of host defense breakdown in sleep deprived rats. Preliminary evidence suggests that bacteria translocating from the gut is one source of infecting pathogens indicated by an increased occurrence of viable bacteria in mesenteric lymph nodes of sleep-deprived rats compared with yoked controls and increase bacteria numbers in the ileum and cecum. Bacterial translocation can occur as a consequence of decreased local immunity, and it is expected to lead to increased pathogen adherence, proliferation, and penetration. Decreased resistance to infectious disease in sleep-deprived rat could occur from a direct effect of sleep deprivation on immune function or an indirect effect through hormonal and metabolic changes, such as low plasma thyroid hormones.

Sleep responses to infectious challenge

There are many central nervous system manifestations of infection that most people are subjectively aware of including loss of appetite, fever, inactivity, social withdrawal and somnolence (Reviewed 4). Indeed, fever and somnolence have been hallmarks of illness throughout medical history. Fever has been extensively studied, in contrast, sleep responses during infections were ignored for over two millennia since the time Hippocrates acknowledged disease-associated sleep responses until the mid 1980’s.

The first efforts to quantify changes in sleep over the course of a microbial infection were published by Toth and Krueger (5). In that study, rabbits were inoculated intravenously with gram positive bacteria (Staphylococci aureus) and sleep was recorded for the next two days. For the first 18-24 hours non-rapid eye movement sleep (NREMS) greatly increased and this was accompanied by an increase in electroencephalographic slow-wave activity (EEG SWA). This latter response is thought to be indicative of a greater intensity of NREMS. This initial increase in NREMS and EEG SWA was followed by decreases in both measures for about another 20-24 hours. Throughout this 48 hour period REMS was inhibited and animals had fevers and the other usual signs of infection, e.g., neutrophilia. In several subsequent experiments rabbits were also infected with other gram positive bacteria, gram negative bacteria, fungi (e.g., Candida albicans) or protozoans (e.g., Trypanosoma brucei). In each case biphase NREMS responses and inhibition of rapid eye movement sleep (REMS) were induced by the infectious agents although the timing of the responses were dependent upon the route by which the microbe was administered, the light/dark cycle and the species of microbe used. For example, NREMS responses to gram negative bacteria after intravenous inoculation tended to be very rapid in onset, the increase in NREMS only lasted 6-8 hours, whereas the period of reduced NREMS was prolonged.

Two findings from this series of experiments emphasize just how much remains to be investigated. In one experiment rabbits were put in either a constant light or a constant dark environment then inoculated. Compared to animals similarly inoculated and kept on a 12:12 hour light/dark cycle, the characteristic initial NREMS increases in animals in constant light was greatly augmented. In contrast, in animals kept in constant dark the later NREMS inhibitory response was greatly enhanced. Reasons for these diametrically opposite responses remain unknown. In another study, the disease commonly called sleeping sickness was studied in rabbits. Rabbits infected with T. brucei brucei exhibit parasitemia about every 7 days during the initial weeks of the disease. These periods of parasitemia are associated with increases in NREMS. However, these increases are superimposed on a longer-term trend of decreasing NREMS over the course of several weeks. Nevertheless, since each episode of parasitemia presents an immune stimulus to the host, the results are consistent with the hypothesis that immune stimulation is correlated with somnolence.

Viral diseases are also associated with changes in sleep. Thus, in humans infected with HIV, but not having AIDS, there is an excess of deep NREMS in the latter half of the night. In contrast, after AIDS develops sleep is greatly disrupted. Rabies viral infections are also associated with sleep changes in animals. However, both HIV and rabies viruses are trophic to the brain and it is thus difficult to distinguish viral-induced tissue damage effects from viral-induced immune effects on sleep. In fact, an early seminal study in sleep research by von Economo used viral-induced brain lesions to localize areas in brain important in sleep regulation. In contrast to these CNS-viral diseases, influenza infections in mice and man are localized to the lung. Sleep in influenza-viral infected mice is greatly altered over many days (Fig. 1). NREMS responses are similar to those observed after bacterial challenge in that there are profound increases in NREMS; these increases however, lasted throughout the 3-day postinoculation period. REMS was inhibited during the 3-day postinoculation period as was wakefulness.

The question of whether sleep responses to infectious challenge are beneficial to host defenses remains unanswered. However, there is a correlation between an animal’s ability to survive an infection and the amount of time spent in NREMS during the first 12 postinoculation hours (Fig. 2).

Research in this area began as an outgrowth of the search for endogenous sleep-promoting substances. One of the somnogenic substances identified in the early 1980’s is a muramyl peptide. Muramyl peptides are well known as the monomeric building blocks of bacterial cell wall peptidoglycan. It was assumed at the time of the isolation of somnogenic muramyl peptides from mammalian tissues that they were derived from bacteria since there are no known synthetic pathways in mammals for some of the components in muramyl peptides, e.g., muramic acid. Subsequently, it was shown, as described above, that sleep loss affects bacterial populations within the gut and the translocation of bacteria across the intestinal wall. Further, mammalian macrophages enzymatically tailor somnogenic muramyl peptides from bacterial cell walls and release them into the surrounding extracellular fluid. Finally, somnogenic muramyl peptides stimulate cytokine production; as described below cytokines seem to be involved in physiological sleep regulation. Collectively, these findings lead to the hypothesis that gut bacteria are affected by sleep and affect sleep; this is truly exciting since it adds yet another dimension to endosymbiotic relationships.

Mechanisms of microbial-induced sleep responses

Infection is also associated with increased growth hormone release. IL-1 also induces growth hormone release and does so via brain-derived growth hormone releasing hormone (GHRH). GHRH is also involved in IL-1-induced sleep; blocking GHRH blocks IL-1-induced NREMS responses. Infection and cytokines also affect several other endocrine systems such as the corticotropin releasing hormone-adrenocorticotropin releasing hormone-glucocorticoid axis. This axis is best characterized as being turned on by stress; many of the individual hormones in this axis inhibit, rather than promote, sleep. It is possible that the upregulation of this axis is responsible for the increased wakefulness which follows microbial-induced sleep responses.

Cytokines and other somnogenic growth factors likely regulate sleep via multiple interactions with neuronal networks (6). The production of some of these growth factors seems to be dependent, in part, on neuronal use. These somnogenic growth factors also are posited to affect neuronal networks in two inseparable ways. First, they directly influence neuronal circuits in several ways including changing the cellular composition of the circuits and their responsiveness to neurotransmitters. Second, these somnogenic growth factors likely are involved in the sculpturing of synaptic populations which, in turn, would affect circuit dynamics. This synaptic sculpturing process could be considered a function of sleep. Thus, sleep mechanisms (circuit dynamics) and sleep function (synaptic sculpturing) are inseparable since they affect each other and depend upon the same growth factors. Regardless of such ideas, it is now gradually becoming clear that sleep is capable of affecting host defense systems and that infectious challenge is associated with robust sleep responses. Such findings will likely add a degree of rationality to what is probably the oldest and most frequently prescribed medicine, sleep.

References

1. Everson, C. A. Sleep deprivation and the immune system. In: Sleep and Biological Rhythms in Health and Sickness, M. R. Pressman and W. C. Orr (Eds.), New York, American Psychological Association, in press.

2. Everson, C. A. Functional consequences of sustained sleep deprivation in the rat. Behav. Brain Res. 69: 43-54, 1995.

3. Everson, C. A. Sustained sleep deprivation impairs host defense. Am. J. Physiol. 265: R1148-R1154, 1993.

4. Krueger, J. M. and J. A. Majde. Microbial products and cytokines in sleep and fever regulation. Crit. Rev. in Immunol. 14: 355-379, 1994.

5. Toth, L. A. and J. M. Krueger. Alteration of sleep in rabbits by Staphylococcus aureus infection. Infect. Immunity 56: 1785-1791, 1988.

6. Krueger, J. M., F. Obál, Jr., M. Opp, L. Toth, L. Johannsen and A. B. Cady. Somnogenic cytokines and models concerning their effects on sleep. Yale J. Biol. Med. 63: 157-172, 1990.

Figure Legends